Not to entirely trash your reasoning Dave, but do you realise that square waves don't occur in analogue signals at all? I think though that Class D amps appear to suffer the same problems that digital does in general being that they essentially digitise the sound first to amplify it.
Jon Atkinson described the Bel Canto as "unhappy" when dealing with high frequencies, particularly with lower impedance loads. A well designed Class A or Class A/B amp is unlikely to rise past 0.05% THD when driving a 2Ohm load at 20kHz. The REF1000 is closer to 0.5%. Regardless, I'm not the type of person who's obsessed with metered measurements. The Class D sound for me has always been unrefined treble that's dark and grainy, with a lack of air around cymbals and other HF instruments. There's something "off" about it. If little 20lb 1000W amplifiers were a free lunch, everybody would be making them. Who would buy a 150lb. space heater? You'd have to be insane. They are not a free lunch, and that hasn't happened.